In a data processing system with multiple compute elements and a shared memory pool, one of the keys to fast and effective processing is the ability to balance the processing load of compute-heavy tasks among the various compute elements. Such accessing and processing must be performed at the multiple compute elements in parallel, simultaneously, in order to maximize the effective processing rate of the system. Needed are systems and methods to achieve load balancing in simultaneous access and processing by multiple compute elements of multiple data sets stored in a shared memory pool.